Foreward
In Part I of this series of articles we discussed the development of time-lapse refraction seismic as a complementary monitoring method. Here we look at some field examples.
The Valhall field. Estimated recoverable reserves from the field at start of production in 1982 were 250 MMbo. Now, after 35 years of production, a total of 1,000 MMboe has been produced out of the original estimated 2.74 Bbo in place. The oil production rate in 2016 averaged ~44,000 boepd, and the field is expected to be producing until beyond 2040. The Valhall ‘life of field seismic’ (LoFS) project has now been running since 2003. (Source: BP Aker)
Southern North Sea
Figure 1: Field data example using LoFS data from Valhall: from survey 8 to survey 1 we observe a clear shift towards shorter offsets for the interpreted amplitude increase close to critical offset (4,000–4,500m). Figure from Zadeh et al., 2011.
The chalk fields in the southern part of the North Sea are good candidates for testing time-lapse refraction sensitivity. One example using the LoFS (Life of Field Seismic) data acquired over the Valhall field confirms this sensitivity (Zadeh et al., 2011). From Figure 1 we observe that the strong amplitude anomaly observed around 4,500m offset is shifted between the first (blue line) and the sixth (red line) LoFS surveys. The offset shift is of the order of 200m, and an additional offset shift is observed for the 8th LoFS survey (black line). This shortening of the critical offset is interpreted to be caused by velocity changes close to the reservoir.
Another potential use of the time-lapse refraction method is as a tool for the early detection of gas leakage or stress changes in the overburden. An example from an underground blow-out that occurred in well 2/4-14 in the North Sea is shown in Figure 2. By studying the refracted wave before and after the blow-out, we clearly observe a time shift for these waves (Zadeh and Landrø, 2011). A more comprehensive analysis is shown in Figure 3, where time shifts are estimated along a 2D profile. We clearly observe that the time shifts increase to a maximum value of around 3–4 ms close to the blow-out well, and decrease away from it, as expected. Different colors in this plot represent different source-receiver offsets used in the time-lapse seismic analysis.
Figure 2: Observed time shift for the refracted wave close to the blow-out well (left) and similar display far away from the well (right). Black traces are before blow-out and red traces after. Typical time shifts within the gas anomaly are of the order of 3-4 ms. (Source: )
Figure 3: Estimated refraction time shifts for three offsets close to the blow-out well (14) – a positive time shift of up to 4 ms is observed. (Source: )
Time-Lapse Refraction: Heavy Oil and Ice
A third example of time-lapse refraction was presented by Hansteen et al., 2010, where time shifts are observed close to a region where steam is injected into a heavy oil reservoir in Alberta, Canada. When the steam is injected, the stiffness of the rock is reduced, leading to a velocity decrease, which in turn leads to a corresponding increase in time shift. This was sucessfully observed in this field example.
A very different approach was used by Hilbich in 2010 where he exploited time-lapse refraction seismic tomography to study seasonal changes in the ground ice in the Alps. He concludes that the method is capable of detecting temporal changes in alpine permafrost and can identify ground ice degradation. A major strength of the method according to Hilbich is the high vertical resolution potential and the ability to discriminate phase changes between frozen and unfrozen material over time. A significant limitation is the low penetration depth due to strong velocity contrasts between the active layer and the permafrost table.
A Short History of Seismic Exploration
1846: Robert Mallett publishes the article ‘On the dynamics of earthquakes’, and receives a grant of £150 to go to Padula in Italy to record and investigate the damages after the earthquake.
1888: August Schmidt uses traveltime versus distance plots to determine seismic velocities.
1899: Cargill Gilston Knott describes and explains propagation, refraction and reflection of seismic waves at subsurface boundaries. He derives reflection coefficients for non-vertical rays which are essentially the same equations that we refer to today as the Zoeppritz equations, which were published 20 years later, in 1919. Knott spent several years in Tokyo and received the Order of the Rising Sun.
Cargill Gilston Knott (1856–1922) was the first to explain refraction.
1910: Andrija Mohorovičić identifies distinct phases of P- and S-waves on traveltime plots from earthquake data. He interpreted them as refractions from a boundary between low and high velocity layers. This boundary, which separates the earth’s crust from the mantle below, is the Moho.
Andrija Mohorovičić (1857–1936) was a Croatian meteorologist and seismologist.
1916: Refraction seismic is used to determine the location of heavy artillery by studying the refracted waves generated by the recoil when the guns were fired. Both the Allies and Germany developed this method, and Ludger Mintrop was especially active in this work; he patented a portable seismograph in 1919.
Ludger Mintrop (1880–1956) developed the portable seismograph and used seismic refraction to map salt structures.
1919–1921: K.C. Karcher carries out the first seismic reflection survey, based on pioneering work done in 1913 by Reginald Fessenden.
1921: Ludger Mintrop founds ‘Seismos Gesellschaft’ with the ambition of using seismic refraction methods to explore and map salt domes acting as traps for hydrocarbons.
1924: The Orchard Salt Dome in Texas is discovered by using refraction seismic and the equipment of Seismos.
1927: Seismic reflection is used routinely for hydrocarbon exploration
Time-Lapse Refraction Radar
Figure 4: Time-lapse refraction radar. Outline of the Grane field (color bar shows depth to the reservoir), which is equipped with a permanent array at the seabed, and if we assume that a seismic source is fired every day at the platform, the refracted signals may be detected at the seabed by the array, and hence it is possible (at least in theory!) to construct a time-lapse refraction radar, sweeping over the field. (Source: Statoil)
Since 2003 several of the major fields offshore Norway have been equipped with permanent receiver arrays at the seabed. The Valhall LoFS-project has been running since then, providing useful data for the operator. At a later date the Ekofisk, Snorre and Grane fields installed similar permanent monitoring systems. The technologies of the systems have been different, but the key concept of semi-continuous monitoring of the field during production has been the same. Typically, these fields are monitored more frequently using this method than when conventional time-lapse seismic surveys are used. A frequency of one to two surveys a year is normal.
Permanent arrays can also be used to monitor micro seismic events, earthquakes, vessel noise, drilling noise and noise from ship traffic. Another potential use of a permanent array is to deploy a simple source firing from the platform (a single gun or a small cluster of airguns may be used for this purpose), and then record the refracted seismic signals from shallow (and if possible, deeper) layers, as illustrated in Figure 4. The time interval between each shot at the platform might be once a month, once a day or even once every hour if one wants to follow a specific event. Under normal conditions, this radar system will not detect any changes; the data from one survey is more or less identical to the one from the previous one apart from noise. A detection system needs to be developed to register abnormal signals when they occur, and how they develop over time. The potential strength of this method is that it is cheap (assuming that the cost of the permanent array is already paid) and the processing or analysis of the data is also fairly straightforward. Weaknesses are related to practical issues regarding firing a seismic airgun from a platform, and to the fact that it is challenging to locate an anomalous event at depth. The x and y coordinates of an event are easily and very precisely determined since the source and the receiver positions are known. Depth can be estimated roughly from seismic modeling. So far, this has not been tested at any field, and therefore there is a chance that refraction radar will remain what it is today: an idea.
The Future of Time-Lapse Refraction
At present time-lapse refraction is not well established as a 4D technique, but the few examples presented so far illustrate that the technique has potential. One weakness is that it is hard to map the time-lapse anomalies precisely in depth. Fortunately, full waveform inversion techniques have developed significantly over the last decade (Virieux and Operto 2009, Sirgue et al., 2009.). A few examples of this technique have also been presented for 4D applications (Routh et al., 2009). Therefore, there is hope that time-lapse refraction methods might develop into a more precise and accurate technique. It is important to underline the fact that if 4D refraction is used as a tool for identifying shallow gas leakage, the critical issue is to detect and locate approximately where the leakage occurs. Detailed and accurate mapping of the 4D anomaly is not crucial.
Further reading
Time-Lapse Refraction Seismic – Part I by Martin Landrø and Lasse Amundsen
A Complementary Monitoring Method? This article appeared in Vol. 14, No. 3
